Major salivary duct carcinomas (MSDCs) often involve the regional lymph nodes (LNs). However, the clinical value of LN parameters in patients with MSDCs is unclear. We aimed to investigate the optimal cut-off points for number of positive LNs (PLNN) and ratio of positive LNs (PLNRs) and their prognostic value in patients with MSDC.
Study Design
Retrospective cohort.
Methods
We retrospectively reviewed relevant data extracted from the Surveillance, Epidemiology, and End Results database on patients with MSDC who had undergone surgery between 2004 and 2016. The optimal PLNN and positive lymph node ratio (PLNR) cut-off points were identified using the X-tile program. Kaplan–Meier and Cox regression analyses were performed to determine prognostic factors.
Results
Overall, 290 patients were enrolled, 57.6% of whom had LN metastases. Advanced T stage in the submandibular gland and unpaired lesions were associated with LN involvement. Positive LNs, late T stage, and submandibular gland location were associated with poor overall survival (OS). The 5-year OS rates of patients with negative and positive LNs were 74.3% and 36.5%, respectively. PLNN > 16 and PLNR > 0.48 were the best cut-off points. The 5-year OS of patients with PLNN ≤ 16 and PLNN > 16 was 42.8% and 15.4%, respectively. The 5-year OS rates were 46.8% for patients with PLNR ≤ 0.48 and 26.3% for patients with PLNR > 0.48. PLNR was a strong prognostic factor for patients with MSDC with LN metastases.
Conclusions
PLNR reflects both the effects of LN dissection and PLNN. Furthermore, its prognostic value in patients with MSDC exceeds that of PLNN. Patients with high PLNRs should be followed closely after surgery.
{"title":"Prognostic significance of ratio of positive lymph nodes in patients with operable major salivary ductal carcinoma","authors":"Di Zhang, Lixi Li","doi":"10.1002/med4.79","DOIUrl":"https://doi.org/10.1002/med4.79","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Background</h3>\u0000 \u0000 <p>Major salivary duct carcinomas (MSDCs) often involve the regional lymph nodes (LNs). However, the clinical value of LN parameters in patients with MSDCs is unclear. We aimed to investigate the optimal cut-off points for number of positive LNs (PLNN) and ratio of positive LNs (PLNRs) and their prognostic value in patients with MSDC.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Study Design</h3>\u0000 \u0000 <p>Retrospective cohort.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Methods</h3>\u0000 \u0000 <p>We retrospectively reviewed relevant data extracted from the Surveillance, Epidemiology, and End Results database on patients with MSDC who had undergone surgery between 2004 and 2016. The optimal PLNN and positive lymph node ratio (PLNR) cut-off points were identified using the X-tile program. Kaplan–Meier and Cox regression analyses were performed to determine prognostic factors.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Results</h3>\u0000 \u0000 <p>Overall, 290 patients were enrolled, 57.6% of whom had LN metastases. Advanced T stage in the submandibular gland and unpaired lesions were associated with LN involvement. Positive LNs, late T stage, and submandibular gland location were associated with poor overall survival (OS). The 5-year OS rates of patients with negative and positive LNs were 74.3% and 36.5%, respectively. PLNN > 16 and PLNR > 0.48 were the best cut-off points. The 5-year OS of patients with PLNN ≤ 16 and PLNN > 16 was 42.8% and 15.4%, respectively. The 5-year OS rates were 46.8% for patients with PLNR ≤ 0.48 and 26.3% for patients with PLNR > 0.48. PLNR was a strong prognostic factor for patients with MSDC with LN metastases.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Conclusions</h3>\u0000 \u0000 <p>PLNR reflects both the effects of LN dissection and PLNN. Furthermore, its prognostic value in patients with MSDC exceeds that of PLNN. Patients with high PLNRs should be followed closely after surgery.</p>\u0000 </section>\u0000 </div>","PeriodicalId":100913,"journal":{"name":"Medicine Advances","volume":"2 3","pages":"291-301"},"PeriodicalIF":0.0,"publicationDate":"2024-09-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1002/med4.79","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142430095","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Mycobacterium tuberculosis, which mainly affects the lungs but can also invade other systems, presents a diagnostic challenge when co-existing with autoimmune disease. We report a case of 31-year-old Asian woman with anti-synthetase syndrome who developed abdominal and subcutaneous lumps during treatment, along with elevated tumor markers, raising suspicion of malignancy. Pathological examination revealed M. tuberculosis infection, changing the diagnosis to multifocal extrapulmonary tuberculosis. This case highlights the need for prompt and detailed evaluations, and the importance of broad differential diagnosis.
{"title":"Anti-synthetase syndrome complicated by multifocal tuberculosis: A thought-provoking differential diagnosis with tumors","authors":"Chang Dai, Yu Jiang, Weidong Lin, Yuhan Chen, Fangzhou Liu, Hai Deng, Yanlin Chen, Haowei Chen, Jieying Wang, Yunxia Lei, Shulin Wu, Yumei Xue","doi":"10.1002/med4.78","DOIUrl":"https://doi.org/10.1002/med4.78","url":null,"abstract":"<p><i>Mycobacterium tuberculosis</i>, which mainly affects the lungs but can also invade other systems, presents a diagnostic challenge when co-existing with autoimmune disease. We report a case of 31-year-old Asian woman with anti-synthetase syndrome who developed abdominal and subcutaneous lumps during treatment, along with elevated tumor markers, raising suspicion of malignancy. Pathological examination revealed <i>M</i>. <i>tuberculosis</i> infection, changing the diagnosis to multifocal extrapulmonary tuberculosis. This case highlights the need for prompt and detailed evaluations, and the importance of broad differential diagnosis.</p>","PeriodicalId":100913,"journal":{"name":"Medicine Advances","volume":"2 3","pages":"312-317"},"PeriodicalIF":0.0,"publicationDate":"2024-09-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1002/med4.78","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142430073","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Clark J. Chen, Vivek K. Bilolikar, Duncan VanNest, James Raphael, Gene Shaffer
Background
This study evaluated the performance of generative artificial intelligence (AI) models on the Orthopaedic In-Training Examination (OITE), an annual exam administered to U.S. orthopaedic residency programs.
Methods
ChatGPT 3.5 and Bing AI GPT 4.0 were evaluated on standardised sets of multiple-choice questions drawn from the American Academy of Orthopaedic Surgeons OITE online question bank spanning 5 years (2018–2022). A total of 1165 questions were posed to each AI system. The performance of both systems was standardised using the latest versions of ChatGPT 3.5 and Bing AI GPT 4.0. Historical data of resident scores taken from the annual OITE technical reports was used as a comparison.
Results
Across the five datasets, ChatGPT 3.5 scored an average of 55.0% on the OITE questions. Bing AI GPT 4.0 scored higher with an average of 80.0%. In comparison, the average performance of orthopaedic residents in national accredited programs was 62.1%. Bing AI GPT 4.0 outperformed ChatGPT 3.5 and Accreditation Council for Graduate Medical Education examinees, and analysis of variance analysis demonstrated p < 0.001 among groups. The best performance was by Bing AI GPT 4.0 on OITE 2020.
Conclusion
Generative AI can provide a logical context across answer responses through its in-depth information searches and citation of resources. This combination presents a convincing argument for the possible uses of AI in medical education as an interactive learning aid.
{"title":"Artificial intelligence in orthopaedic education: A comparative analysis of ChatGPT and Bing AI's Orthopaedic In-Training Examination performance","authors":"Clark J. Chen, Vivek K. Bilolikar, Duncan VanNest, James Raphael, Gene Shaffer","doi":"10.1002/med4.77","DOIUrl":"https://doi.org/10.1002/med4.77","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Background</h3>\u0000 \u0000 <p>This study evaluated the performance of generative artificial intelligence (AI) models on the Orthopaedic In-Training Examination (OITE), an annual exam administered to U.S. orthopaedic residency programs.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Methods</h3>\u0000 \u0000 <p>ChatGPT 3.5 and Bing AI GPT 4.0 were evaluated on standardised sets of multiple-choice questions drawn from the American Academy of Orthopaedic Surgeons OITE online question bank spanning 5 years (2018–2022). A total of 1165 questions were posed to each AI system. The performance of both systems was standardised using the latest versions of ChatGPT 3.5 and Bing AI GPT 4.0. Historical data of resident scores taken from the annual OITE technical reports was used as a comparison.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Results</h3>\u0000 \u0000 <p>Across the five datasets, ChatGPT 3.5 scored an average of 55.0% on the OITE questions. Bing AI GPT 4.0 scored higher with an average of 80.0%. In comparison, the average performance of orthopaedic residents in national accredited programs was 62.1%. Bing AI GPT 4.0 outperformed ChatGPT 3.5 and Accreditation Council for Graduate Medical Education examinees, and analysis of variance analysis demonstrated <i>p</i> < 0.001 among groups. The best performance was by Bing AI GPT 4.0 on OITE 2020.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Conclusion</h3>\u0000 \u0000 <p>Generative AI can provide a logical context across answer responses through its in-depth information searches and citation of resources. This combination presents a convincing argument for the possible uses of AI in medical education as an interactive learning aid.</p>\u0000 </section>\u0000 </div>","PeriodicalId":100913,"journal":{"name":"Medicine Advances","volume":"2 3","pages":"284-290"},"PeriodicalIF":0.0,"publicationDate":"2024-09-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1002/med4.77","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142430061","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Patient navigators, whose value has become increasingly apparent, still face significant challenges, including a lack of support, funding, and recognition. These challenges have been exacerbated in the wake of COVID-19 pandemic.
Methods
This study explored the potential use of artificial intelligence (AI) in patient navigation. Data were collected through structured surveys and individual interviews with patient navigators from a variety of institutions and professional backgrounds. The data were analyzed to understand the current state of patient navigation, identify existing gaps, and suggest best practices for the future.
Results
The findings showed that patient navigators (a) have diverse backgrounds and responsibilities, (b) lack technology support for their work, (c) are at risk for burnout, with the extent varying based on the level of technical support received, and (d) report significant overlap between current barriers and those that could potentially be addressed with AI-driven technologies.
Conclusion
A novel intervention, that is enabled by AI and other technologies and tailored to individual needs, has the potential to reduce burnout, increase capacity, and help ensure the sustainability of patient navigation and other areas of healthcare. By addressing the specific needs of individual patients, this type of intervention could help improve the overall effectiveness of patient navigation and support the long-term sustainability of the role.
{"title":"Toward bridging gaps in patient navigation: A study on the adoption of artificial intelligence technologies","authors":"Fenghao Chen, Tu Lan, Jie Liang, Ronghui Zhang","doi":"10.1002/med4.76","DOIUrl":"https://doi.org/10.1002/med4.76","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Background</h3>\u0000 \u0000 <p>Patient navigators, whose value has become increasingly apparent, still face significant challenges, including a lack of support, funding, and recognition. These challenges have been exacerbated in the wake of COVID-19 pandemic.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Methods</h3>\u0000 \u0000 <p>This study explored the potential use of artificial intelligence (AI) in patient navigation. Data were collected through structured surveys and individual interviews with patient navigators from a variety of institutions and professional backgrounds. The data were analyzed to understand the current state of patient navigation, identify existing gaps, and suggest best practices for the future.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Results</h3>\u0000 \u0000 <p>The findings showed that patient navigators (a) have diverse backgrounds and responsibilities, (b) lack technology support for their work, (c) are at risk for burnout, with the extent varying based on the level of technical support received, and (d) report significant overlap between current barriers and those that could potentially be addressed with AI-driven technologies.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Conclusion</h3>\u0000 \u0000 <p>A novel intervention, that is enabled by AI and other technologies and tailored to individual needs, has the potential to reduce burnout, increase capacity, and help ensure the sustainability of patient navigation and other areas of healthcare. By addressing the specific needs of individual patients, this type of intervention could help improve the overall effectiveness of patient navigation and support the long-term sustainability of the role.</p>\u0000 </section>\u0000 </div>","PeriodicalId":100913,"journal":{"name":"Medicine Advances","volume":"2 3","pages":"274-283"},"PeriodicalIF":0.0,"publicationDate":"2024-09-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1002/med4.76","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142429824","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Clear cell renal cell carcinoma (ccRCC) is one of the most common urological diseases and the most common subtype of renal cell carcinoma. Nuclear mitochondria-related genes (MTRGs) play an essential role in cancer, but their effect on ccRCC has not been clarified. This work aimed to investigate the role of nuclear MTRGs in ccRCC.
Methods
We collected nuclear MTRGs from the MITOMAP database and obtained the ccRCC profile from the TCGA database. Gene expression validation came from the GEO database. The ccRCC subtypes were determined by unsupervised clustering analysis based on the nuclear MTRGs. Immune scores were computed using the EPIC algorithm. The drug sensitivity scores were calculated using GDSC resources. A nomogram explored the diagnostic value of the nuclear MTRGs.
Results
In total, 11 nuclear MTRGs were identified as both related to prognosis and differentially expressed in ccRCC, showing a significant positive correlation with CD274 expression. We determined two subtypes of ccRCC based on these genes and found remarkable differences in survival status, immune infiltration, mutation landscape, and drug sensitivity between the two subtypes. The high-MTRG group had a better prognosis and a lower tumor stage than the low-MTRG group. Immune checkpoint blockade therapy was more effective for the high-MTRG group. A nomogram based on the nuclear MTRGs concluded that patients with a higher score had poorer survival. SUCLA2 (succinate-CoA ligase ADP-forming subunit beta) was identified as the hub gene linked to ccRCC. High SUCLA2 expression showed a correlation with better survival and a negative correlation with the tumor mutation burden in ccRCC. Pan-cancer analysis revealed wide-ranging roles for SUCLA2 across human tumors.
Conclusions
Nuclear MTRGs play vital roles in determining the molecular features, immune infiltration, and drug sensitivity of ccRCC. High levels of nuclear MTRGs may indicate a better prognosis for patients with ccRCC. SUCLA2 is a representative nuclear MTRG and may serve as a protective biomarker in ccRCC. Our study provides therapeutic guidance and potential biomarkers for ccRCC patients, and contributes to the advancement of precision medicine.
{"title":"Comprehensive analyses of nuclear mitochondria-related genes in the molecular features, immune infiltration, and drug sensitivity of clear cell renal cell carcinoma","authors":"Yuchen Zhang, Huake Cao, Feixiang Yang, Xiaofeng Wang, Zhihao Xu, Miao Cheng, Shuqi Yang, Xuefeng Tian, Ning Zhang, Yinyin Xie","doi":"10.1002/med4.72","DOIUrl":"https://doi.org/10.1002/med4.72","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Background</h3>\u0000 \u0000 <p>Clear cell renal cell carcinoma (ccRCC) is one of the most common urological diseases and the most common subtype of renal cell carcinoma. Nuclear mitochondria-related genes (MTRGs) play an essential role in cancer, but their effect on ccRCC has not been clarified. This work aimed to investigate the role of nuclear MTRGs in ccRCC.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Methods</h3>\u0000 \u0000 <p>We collected nuclear MTRGs from the MITOMAP database and obtained the ccRCC profile from the TCGA database. Gene expression validation came from the GEO database. The ccRCC subtypes were determined by unsupervised clustering analysis based on the nuclear MTRGs. Immune scores were computed using the EPIC algorithm. The drug sensitivity scores were calculated using GDSC resources. A nomogram explored the diagnostic value of the nuclear MTRGs.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Results</h3>\u0000 \u0000 <p>In total, 11 nuclear MTRGs were identified as both related to prognosis and differentially expressed in ccRCC, showing a significant positive correlation with <i>CD274</i> expression. We determined two subtypes of ccRCC based on these genes and found remarkable differences in survival status, immune infiltration, mutation landscape, and drug sensitivity between the two subtypes. The high-MTRG group had a better prognosis and a lower tumor stage than the low-MTRG group. Immune checkpoint blockade therapy was more effective for the high-MTRG group. A nomogram based on the nuclear MTRGs concluded that patients with a higher score had poorer survival. <i>SUCLA2</i> (succinate-CoA ligase ADP-forming subunit beta) was identified as the hub gene linked to ccRCC. High <i>SUCLA2</i> expression showed a correlation with better survival and a negative correlation with the tumor mutation burden in ccRCC. Pan-cancer analysis revealed wide-ranging roles for <i>SUCLA2</i> across human tumors.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Conclusions</h3>\u0000 \u0000 <p>Nuclear MTRGs play vital roles in determining the molecular features, immune infiltration, and drug sensitivity of ccRCC. High levels of nuclear MTRGs may indicate a better prognosis for patients with ccRCC. <i>SUCLA2</i> is a representative nuclear MTRG and may serve as a protective biomarker in ccRCC. Our study provides therapeutic guidance and potential biomarkers for ccRCC patients, and contributes to the advancement of precision medicine.</p>\u0000 </section>\u0000 </div>","PeriodicalId":100913,"journal":{"name":"Medicine Advances","volume":"2 3","pages":"238-253"},"PeriodicalIF":0.0,"publicationDate":"2024-09-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1002/med4.72","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142429287","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Han Yuan, Kunyu Yu, Feng Xie, Mingxuan Liu, Shenghuan Sun
Machine learning (ML) has achieved substantial success in performing healthcare tasks in which the configuration of every part of the ML pipeline relies heavily on technical knowledge. To help professionals with borderline expertise to better use ML techniques, Automated ML (AutoML) has emerged as a prospective solution. However, most models generated by AutoML are black boxes that are challenging to comprehend and deploy in healthcare settings. We conducted a systematic review to examine AutoML with interpretation systems for healthcare. We searched four databases (MEDLINE, EMBASE, Web of Science, and Scopus) complemented with seven prestigious ML conferences (AAAI, ACL, ICLR, ICML, IJCAI, KDD, and NeurIPS) that reported AutoML with interpretation for healthcare before September 1, 2023. We included 118 articles related to AutoML with interpretation in healthcare. First, we illustrated AutoML techniques used in the included publications, including automated data preparation, automated feature engineering, and automated model development, accompanied by a real-world case study to demonstrate the advantages of AutoML over classic ML. Then, we summarized interpretation methods: feature interaction and importance, data dimensionality reduction, intrinsically interpretable models, and knowledge distillation and rule extraction. Finally, we detailed how AutoML with interpretation has been used for six major data types: image, free text, tabular data, signal, genomic sequences, and multi-modality. To some extent, AutoML with interpretation provides effortless development and improves users' trust in ML in healthcare settings. In future studies, researchers should explore automated data preparation, seamless integration of automation and interpretation, compatibility with multi-modality, and utilization of foundation models.
机器学习（ML）在执行医疗保健任务方面取得了巨大成功，其中 ML 管道每个部分的配置都严重依赖于技术知识。为了帮助专业人员更好地使用 ML 技术，自动 ML（AutoML）已成为一种前瞻性解决方案。然而，AutoML 生成的大多数模型都是黑盒子，在医疗环境中理解和部署具有挑战性。我们进行了一项系统性综述，研究了用于医疗保健的带有解释系统的 AutoML。我们检索了四个数据库（MEDLINE、EMBASE、Web of Science 和 Scopus）以及七个著名的 ML 会议（AAAI、ACL、ICLR、ICML、IJCAI、KDD 和 NeurIPS），这些会议都报道了 2023 年 9 月 1 日之前用于医疗保健的带解释的 AutoML。我们收录了 118 篇与医疗保健领域的 AutoML 与解释相关的文章。首先，我们介绍了所收录文章中使用的 AutoML 技术，包括自动数据准备、自动特征工程和自动模型开发，并通过实际案例研究展示了 AutoML 相对于传统 ML 的优势。然后，我们总结了解释方法：特征交互和重要性、数据降维、内在可解释模型以及知识提炼和规则提取。最后，我们详细介绍了带有解释功能的 AutoML 如何用于六种主要数据类型：图像、自由文本、表格数据、信号、基因组序列和多模态。在某种程度上，带有解释功能的 AutoML 可以毫不费力地进行开发，并提高用户对医疗保健领域人工智能的信任度。在未来的研究中，研究人员应探索自动化数据准备、自动化与解释的无缝集成、与多模态的兼容性以及基础模型的利用。
{"title":"Automated machine learning with interpretation: A systematic review of methodologies and applications in healthcare","authors":"Han Yuan, Kunyu Yu, Feng Xie, Mingxuan Liu, Shenghuan Sun","doi":"10.1002/med4.75","DOIUrl":"https://doi.org/10.1002/med4.75","url":null,"abstract":"<p>Machine learning (ML) has achieved substantial success in performing healthcare tasks in which the configuration of every part of the ML pipeline relies heavily on technical knowledge. To help professionals with borderline expertise to better use ML techniques, Automated ML (AutoML) has emerged as a prospective solution. However, most models generated by AutoML are black boxes that are challenging to comprehend and deploy in healthcare settings. We conducted a systematic review to examine AutoML with interpretation systems for healthcare. We searched four databases (MEDLINE, EMBASE, Web of Science, and Scopus) complemented with seven prestigious ML conferences (AAAI, ACL, ICLR, ICML, IJCAI, KDD, and NeurIPS) that reported AutoML with interpretation for healthcare before September 1, 2023. We included 118 articles related to AutoML with interpretation in healthcare. First, we illustrated AutoML techniques used in the included publications, including automated data preparation, automated feature engineering, and automated model development, accompanied by a real-world case study to demonstrate the advantages of AutoML over classic ML. Then, we summarized interpretation methods: feature interaction and importance, data dimensionality reduction, intrinsically interpretable models, and knowledge distillation and rule extraction. Finally, we detailed how AutoML with interpretation has been used for six major data types: image, free text, tabular data, signal, genomic sequences, and multi-modality. To some extent, AutoML with interpretation provides effortless development and improves users' trust in ML in healthcare settings. In future studies, researchers should explore automated data preparation, seamless integration of automation and interpretation, compatibility with multi-modality, and utilization of foundation models.</p>","PeriodicalId":100913,"journal":{"name":"Medicine Advances","volume":"2 3","pages":"205-237"},"PeriodicalIF":0.0,"publicationDate":"2024-08-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1002/med4.75","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142430183","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The high transmissibility of the Omicron variants of severe acute respiratory syndrome coronavirus 2 continues to impose a significant burden on public health systems worldwide. Continuous mutations in the virus challenge the efficacy of vaccines and prior immunity from other variants, posing a severe threat to human health. Therefore, there is an urgent need to predict mutations of the Omicron variants and identify key factors influencing their spread. This study investigated critical amino acid mutations of the Omicron variants using epidemiological data and the mechanisms underlying their transmission. The aim was to understand the key factors driving the spread of the Omicron variants and provide insights for effective control and prevention strategies.
Methods
A total of 488,646 Omicron cases recorded between December 2021 and February 2023 were analyzed using a sliding time window and the Epi Score model, which has high accuracy for the identification of mutation sites. MutPred2, PolyPhen2 and VarSite tools were used to predict future mutations and identify factors driving the prevalence of the Omicron variants.
Results
Epi Scores showed fluctuating patterns at mutation sites, highlighting N969K, Y505H, N764K, T478K, and S371F mutations on the spike protein as significant for future prevention efforts. The spread of the Omicron variants was linked to changes in the viral entry pathway and improved angiotensin-converting enzyme 2 (ACE2) binding and immune evasion tactics.
Conclusions
The findings of this study reveal trends in the evolution of the Omicron variants, including altered cell entry, increased affinity for ACE2, and evasion of the immune system. These factors are critical for understanding the global spread of the Omicron variants and developing effective control strategies.
{"title":"Internal drivers of the global pandemic of the Omicron variants of SARS-CoV-2","authors":"Yilin Niu, Wuxiu Quan, Yunyi Li, Ziying Chen, Yunfei Gao","doi":"10.1002/med4.74","DOIUrl":"https://doi.org/10.1002/med4.74","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Background</h3>\u0000 \u0000 <p>The high transmissibility of the Omicron variants of severe acute respiratory syndrome coronavirus 2 continues to impose a significant burden on public health systems worldwide. Continuous mutations in the virus challenge the efficacy of vaccines and prior immunity from other variants, posing a severe threat to human health. Therefore, there is an urgent need to predict mutations of the Omicron variants and identify key factors influencing their spread. This study investigated critical amino acid mutations of the Omicron variants using epidemiological data and the mechanisms underlying their transmission. The aim was to understand the key factors driving the spread of the Omicron variants and provide insights for effective control and prevention strategies.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Methods</h3>\u0000 \u0000 <p>A total of 488,646 Omicron cases recorded between December 2021 and February 2023 were analyzed using a sliding time window and the Epi Score model, which has high accuracy for the identification of mutation sites. MutPred2, PolyPhen2 and VarSite tools were used to predict future mutations and identify factors driving the prevalence of the Omicron variants.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Results</h3>\u0000 \u0000 <p>Epi Scores showed fluctuating patterns at mutation sites, highlighting N969K, Y505H, N764K, T478K, and S371F mutations on the spike protein as significant for future prevention efforts. The spread of the Omicron variants was linked to changes in the viral entry pathway and improved angiotensin-converting enzyme 2 (ACE2) binding and immune evasion tactics.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Conclusions</h3>\u0000 \u0000 <p>The findings of this study reveal trends in the evolution of the Omicron variants, including altered cell entry, increased affinity for ACE2, and evasion of the immune system. These factors are critical for understanding the global spread of the Omicron variants and developing effective control strategies.</p>\u0000 </section>\u0000 </div>","PeriodicalId":100913,"journal":{"name":"Medicine Advances","volume":"2 3","pages":"262-273"},"PeriodicalIF":0.0,"publicationDate":"2024-08-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1002/med4.74","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142430306","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
This article performs a literature search to determine the current progress, identifies research gaps, and highlights future opportunities of human-in-the-loop across the machine learning lifecycle, including data preparation, feature engineering, model development, and model deployment.
{"title":"Human-in-the-loop machine learning for healthcare: Current progress and future opportunities in electronic health records","authors":"Han Yuan, Lican Kang, Yong Li, Zhenqian Fan","doi":"10.1002/med4.70","DOIUrl":"https://doi.org/10.1002/med4.70","url":null,"abstract":"<p>This article performs a literature search to determine the current progress, identifies research gaps, and highlights future opportunities of human-in-the-loop across the machine learning lifecycle, including data preparation, feature engineering, model development, and model deployment.\u0000 <figure>\u0000 <div><picture>\u0000 <source></source></picture><p></p>\u0000 </div>\u0000 </figure></p>","PeriodicalId":100913,"journal":{"name":"Medicine Advances","volume":"2 3","pages":"318-322"},"PeriodicalIF":0.0,"publicationDate":"2024-08-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1002/med4.70","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142430176","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Agitated saline contrast echocardiography, essential for patent foramen ovale detection, typically uses a 10 mL saline protocol, which may lead to operator fatigue and compromise diagnostic accuracy. This study aimed to identify an optimal saline volume that balances diagnostic efficacy and minimizes operator fatigue.
Methods
Fifty-three patients with suspected patent foramen ovale underwent echocardiography using three saline protocols: 10 mL (9 mL saline + 1 mL air), 7 mL (6.4 mL saline + 0.6 mL air), and 5 mL (4.6 mL saline + 0.4 mL air). The protocols were evaluated for their impact on microbubble signal intensity in the right atrium, right-to-left shunt grading, and operator fatigue, as assessed by the Borg CR10 scale.
Results
The signal intensities were comparable between the 10 mL and 7 mL protocols, while the 5 mL protocol showed significantly lower signal intensities at all time points (p < 0.001). The 7 mL protocol achieved 100% right-to-left shunt grading concordance with the 10 mL protocol, while the 5 mL protocol showed decreased accuracy and recall. Operator fatigue significantly decreased with the reduction in saline volume (p < 0.001), and the 7 mL protocol demonstrated optimal diagnostic accuracy and reduced fatigue.
Conclusion
A 7 mL agitated saline volume is recommended for contrast echocardiography to maintain diagnostic accuracy while significantly reducing operator fatigue, offering a practical balance between diagnostic precision and operator comfort.
背景 搅拌生理盐水造影超声心动图是检测卵圆孔未闭的必要方法，通常使用 10 mL 生理盐水，这可能会导致操作者疲劳并影响诊断准确性。本研究旨在找出既能兼顾诊断效果又能最大程度减轻操作者疲劳的最佳生理盐水量。 方法 53 名疑似卵圆孔未闭患者使用三种生理盐水方案接受超声心动图检查：10 mL（9 mL 生理盐水 + 1 mL 空气）、7 mL（6.4 mL 生理盐水 + 0.6 mL 空气）和 5 mL（4.6 mL 生理盐水 + 0.4 mL 空气）。根据 Borg CR10 量表评估了这些方案对右心房微泡信号强度、右向左分流分级和操作者疲劳度的影响。 结果 10 mL 和 7 mL 方案的信号强度相当，而 5 mL 方案在所有时间点的信号强度都明显较低（p < 0.001）。7 mL 方案与 10 mL 方案的右向左分流分级一致性达到 100%，而 5 mL 方案的准确性和召回率有所下降。随着生理盐水容量的减少，操作员的疲劳度明显降低（p < 0.001），7 mL 方案显示出最佳的诊断准确性并减少了疲劳度。 结论 建议在造影超声心动图检查中使用 7 毫升的搅拌生理盐水，以保持诊断准确性，同时显著减轻操作者的疲劳，在诊断准确性和操作舒适性之间取得切实可行的平衡。
{"title":"Optimizing agitated saline volume for contrast echocardiography: Balancing diagnostic performance and operator fatigue","authors":"Hezhi Li, Yanying Huang, Dongling Luo, Song Wen, Hongwen Fei, Caojin Zhang","doi":"10.1002/med4.73","DOIUrl":"https://doi.org/10.1002/med4.73","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Background</h3>\u0000 \u0000 <p>Agitated saline contrast echocardiography, essential for patent foramen ovale detection, typically uses a 10 mL saline protocol, which may lead to operator fatigue and compromise diagnostic accuracy. This study aimed to identify an optimal saline volume that balances diagnostic efficacy and minimizes operator fatigue.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Methods</h3>\u0000 \u0000 <p>Fifty-three patients with suspected patent foramen ovale underwent echocardiography using three saline protocols: 10 mL (9 mL saline + 1 mL air), 7 mL (6.4 mL saline + 0.6 mL air), and 5 mL (4.6 mL saline + 0.4 mL air). The protocols were evaluated for their impact on microbubble signal intensity in the right atrium, right-to-left shunt grading, and operator fatigue, as assessed by the Borg CR10 scale.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Results</h3>\u0000 \u0000 <p>The signal intensities were comparable between the 10 mL and 7 mL protocols, while the 5 mL protocol showed significantly lower signal intensities at all time points (<i>p</i> < 0.001). The 7 mL protocol achieved 100% right-to-left shunt grading concordance with the 10 mL protocol, while the 5 mL protocol showed decreased accuracy and recall. Operator fatigue significantly decreased with the reduction in saline volume (<i>p</i> < 0.001), and the 7 mL protocol demonstrated optimal diagnostic accuracy and reduced fatigue.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Conclusion</h3>\u0000 \u0000 <p>A 7 mL agitated saline volume is recommended for contrast echocardiography to maintain diagnostic accuracy while significantly reducing operator fatigue, offering a practical balance between diagnostic precision and operator comfort.</p>\u0000 </section>\u0000 </div>","PeriodicalId":100913,"journal":{"name":"Medicine Advances","volume":"2 3","pages":"254-261"},"PeriodicalIF":0.0,"publicationDate":"2024-08-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1002/med4.73","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142430104","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Ismail Mohammad Adam, Elfatih Yousif Abdelrahim, Wael Mohialddin Ahmed Doush
Background
Histopathological patterns of periampullary cancers (PAC) provide valuable information for clinicians in the selection of effective postoperative therapies and serve as a prognostic marker for patients. Therefore, this study was conducted to determine the surgical management and histopathological patterns of PAC necessary to achieve excellent surgical outcomes.
Methods
This prospective study was conducted from October 2021 to September 2023 at Ibn Sina Specialized Hospital, Khartoum, Sudan. Eighty patients diagnosed with resectable PAC were enrolled.
Results
PAC was more prevalent in male patients (52.5%), with a male: female ratio of 1.1:1.0. The peak incidence occurred between 46 and 60 years, representing 38.75% of patients. The most significant reported comorbidity was cigarette smoking (36.25% of patients). Scleral icterus was found in 47.5% of patients. Pancreatic head cancer was found in 55% of patients. Histopathological analysis revealed adenocarcinoma in 82.5% of patients. R0 resection (tumor-free margins) was the most common resection category (87.5% of patients). Lymphovascular invasion was detected in 37.5% of patients. Malignant lymph node involvement (1–3 nodes) was found 47.5% of patients. Perineural invasion was found in 55%. Finally, the most common histopathological differentiation of PAC was well-differentiated (Grade I) in 67.5% of patients. The mortality rate was 0%.
Conclusions
Updated knowledge regarding the management modalities of PAC is a prerequisite for accurate histopathological assessment and good surgical outcomes. Furthermore, histopathological analysis helps clinicians to select effective postoperative therapies for greater improvement in survival outcomes.
{"title":"Surgical management and histopathological patterns of periampullary cancers in Sudanese patients: A single-center prospective study","authors":"Ismail Mohammad Adam, Elfatih Yousif Abdelrahim, Wael Mohialddin Ahmed Doush","doi":"10.1002/med4.71","DOIUrl":"https://doi.org/10.1002/med4.71","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Background</h3>\u0000 \u0000 <p>Histopathological patterns of periampullary cancers (PAC) provide valuable information for clinicians in the selection of effective postoperative therapies and serve as a prognostic marker for patients. Therefore, this study was conducted to determine the surgical management and histopathological patterns of PAC necessary to achieve excellent surgical outcomes.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Methods</h3>\u0000 \u0000 <p>This prospective study was conducted from October 2021 to September 2023 at Ibn Sina Specialized Hospital, Khartoum, Sudan. Eighty patients diagnosed with resectable PAC were enrolled.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Results</h3>\u0000 \u0000 <p>PAC was more prevalent in male patients (52.5%), with a male: female ratio of 1.1:1.0. The peak incidence occurred between 46 and 60 years, representing 38.75% of patients. The most significant reported comorbidity was cigarette smoking (36.25% of patients). Scleral icterus was found in 47.5% of patients. Pancreatic head cancer was found in 55% of patients. Histopathological analysis revealed adenocarcinoma in 82.5% of patients. R0 resection (tumor-free margins) was the most common resection category (87.5% of patients). Lymphovascular invasion was detected in 37.5% of patients. Malignant lymph node involvement (1–3 nodes) was found 47.5% of patients. Perineural invasion was found in 55%. Finally, the most common histopathological differentiation of PAC was well-differentiated (Grade I) in 67.5% of patients. The mortality rate was 0%.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Conclusions</h3>\u0000 \u0000 <p>Updated knowledge regarding the management modalities of PAC is a prerequisite for accurate histopathological assessment and good surgical outcomes. Furthermore, histopathological analysis helps clinicians to select effective postoperative therapies for greater improvement in survival outcomes.</p>\u0000 </section>\u0000 </div>","PeriodicalId":100913,"journal":{"name":"Medicine Advances","volume":"2 3","pages":"302-311"},"PeriodicalIF":0.0,"publicationDate":"2024-08-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1002/med4.71","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142430123","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号